The purpose of this clinical study is to examine the role of vasopressin and norepinephrine in the treatment of septic shock and their impact on survival. Norepinephrine, a vasopressor, is the most commonly used clinical agent to reverse the lethal hypotension in patients with septic shock. A new therapy for sepsis has been the use of vasopressin to treat the hypotension associated with sepsis. Small studies have examined vasopressin in conjunction with norepinephrine and suggest that their use together may lower norepinephrine requirements in septic patients. High doses of either vasopressin or norepinephrine can lead to negative outcomes including decreased end organ perfusion and decreased cardiac output. It is believed that by combining these two drugs, there will be less vasoconstriction related organ injury and an increased benefit in survival rate. However, no clinical or animal study has examined the effect of norepinephrine, vasopressin or a combination on survival rate, blood pressure and organ injury. This is, in part, due to the lethality of the disease where withholding such therapies would be impossible. Of particular concern is that it is entirely possible that some of these vasopressors can improve blood pressure but adversely affect survival. L-NMMA, a vasopressor, increased blood pressure but worsened outcome in this model of sepsis. Unfortunately these results were confirmed in humans with septic shock. Previously, in a pilot study, we determined doses of each drug used alone that raised blood pressure but did not adversely affect survival rate. This study will examine the role of vasopressin and norepinephrine used alone and together in treating sepsis in the canine model. We will for the first time in any model of sepsis determine the effects of vasopressin and norepinephrine on blood pressure, cardiac output and survival. The results of this study suggest that the combination treatment of vasopressin and norepinphrine was more beneficial but due to early mortality further experiments will have to be performed to provide sufficient statistical power.